## Screenshot: Task Instructions for Crowdsourcing HIT
### Overview
This image is a screenshot of a web-based instruction page for a Human Intelligence Task (HIT). The document outlines a two-part "detective thinking" exercise where the participant is required to analyze an image to identify "observable clues" and then derive "indications" (non-obvious information) from those clues. The page includes specific rules for formatting, content, and constraints to ensure high-quality data collection.
### Components/Structure
The document is structured as a vertical, top-down instructional guide:
* **Header:** A collapsible UI element labeled "Instructions (click to expand/collapse)".
* **Main Body:** Contains the core task description, procedural steps (Part 1 & 2), bonus opportunities, and a numbered list of rules.
* **Footer:** Two additional collapsible UI elements labeled "How to Pick Good Clues/Indications (click to expand/collapse)" and "Examples (click to expand/collapse)".
### Detailed Analysis
#### Header
* **Text:** "Instructions (click to expand/collapse)"
#### Main Body
* **Opening:** "Thanks for participating in this HIT!"
* **Task Description:** "Your task: In this task, we are asking you to put on your detective thinking cap. Given an image, find observable clues that might indicate information about a person, situation, or setting that may not be necessarily obvious in the image (we will call this indication). You will do this in two steps."
* **PART 1: Examine the image and find 3 observable clues.**
* *Definition:* "An observable clue MUST be something in the picture (e.g., an open algebra math workbook)"
* *Step 1:* "Choose observation number from the drop down box (1 is already chosen for you) and write down your clues you observed in the field to the right (What you write here will be transferred over to the PART 2)."
* *Step 2:* "Draw bounding boxes for the clues (you may draw multiple if there are multiple things you observed)."
* *Step 3:* "Repeat steps 1&2 for all the observations you want to make. Then, move to Part 2 to provide indications for each of the clues you provided."
* **PART 2: For each observable clue, provide a indication.**
* *Definition:* "An indication is a bit of non obvious information about what the clue means to you (e.g., an open algebra math workbook might indicate there might be a high school student who was just studying)"
* *Step 1:* "Write down the indications."
* *Step 2:* "Rate how likely for the indications to be true given the clue"
* "**certain**: It's obvious or i'm very much certain what I said is true (I'm totally willing to bet on it!)."
* "**likely**: It is likely or probable that what I said is true (both moderate and strong likelihood uncertainties belong here)."
* "**possible**: It's in the realm of possibly but it's an educated guess at best."
* *Note:* "We aren't looking for a particular distribution in the ratings nor do we value one rating over another. If you turn in all "possible"s for an image, for example, that's just as acceptable as turning in one of each!"
* **Bonus:** "Five examples are given below the instruction panel. Bonus opportunity: you can provide up to 2 additional clues/indication sets for bonus pay."
* **Rules:**
1. **For observable clues:**
* "Write a noun phrase: 'the book', 'gray skies', 'a group of people'"
* "When possible, please specify details relevant as to where the object, entity, or thing:"
* "'the book' -> 'the book under the table'"
* "'buttons' -> 'buttons on the man's shirt'"
* "'a group of people' -> 'a group of people in the pool'"
* "'a painting' -> 'a painting hanging on the wall'"
* "'a dog' -> 'a dog following a person'"
* "You can provide similar observable clues multiple times, but please tailor your clue to the observation made (see Example 2)."
* "When bounding the clues, please remember: the boxes do not have to be perfect! 1-3 items in a picture is plenty for bounding. Do not spend too much time on this step!"
2. **For indications:**
* "Write in complete sentences"
* "Make the indications realistic"
* "Please DO NOT write indications that contradict each other. For example, indications like 'this is a gathering of family members' and 'this is a work event' cannot both be true. These are contradictions of each other. Please AVOID these."
3. "At this time, we are NOT interested in in plain descriptions of what's going on, what people are doing, and what the people are thinking. Please see 'How to Pick Good Clues/Indications' for further detail."
4. "Please use weather related observations if it's a salient aspect of the image or you have nothing else you can talk about. Please use weather as the last resort. Example 4 observation 2 is an example of a weather observation."
5. "Please avoid gendered pronouns like 'he', 'she', 'him' or 'her'. If you desire, you can use 'they'."
6. "Read through example and how-to sections below!"
#### Footer
* "How to Pick Good Clues/Indications (click to expand/collapse)"
* "Examples (click to expand/collapse)"
### Key Observations
* **Constraint-Heavy:** The instructions are highly specific, likely to ensure the data collected is consistent and usable for training machine learning models or conducting psychological research.
* **Inferential Focus:** The task explicitly discourages "plain descriptions" (Rule 3), pushing the user toward inferential reasoning (e.g., moving from "a book" to "a student studying").
* **Neutrality:** The requirement to avoid gendered pronouns (Rule 5) suggests a desire for gender-neutral data or a reduction in bias.
* **Quality Control:** The inclusion of a rating system (certain/likely/possible) allows the researchers to weigh the confidence of the annotator's inferences.
### Interpretation
This document is a set of guidelines for a crowdsourced data annotation task. The goal is to bridge the gap between **Computer Vision** (identifying objects/clues) and **Natural Language Processing/Reasoning** (inferring context/indications).
The task is designed to extract "latent" information from images—information that is not explicitly stated but is logically implied by the visual evidence. By forcing participants to provide "indications" rather than just descriptions, the task creator is likely building a dataset for training AI models to perform "common sense reasoning" or "social inference" based on visual inputs. The strict rules regarding contradictions and gendered pronouns indicate a high standard for data cleanliness and bias mitigation.